By Millie Muroi, Economics Writer
ChatGPT is not the answer to Australia’s productivity problem. At least, not yet.
But I asked ChatGPT what its chances were of productivity improving in Australia – if it was a betting man. The answer? 70 to 80 per cent.
Productivity growth excites economics nerds, like those at the Reserve Bank ... and just about no one else. But it matters for everything from your mortgage to the prices you pay at shops and the quality of your life.
Why? Because productivity growth means being able to make more with what we have, which is the best solution to the biggest economic issue of our time: inflation.
After all, there are two sides to this inflation problem: too much demand and too little supply. Instead of the Reserve Bank beating down our appetite for goods and services through ramping up interest rates, wouldn’t it be nice if businesses could simply produce more with the workers and equipment they already had, therefore keeping prices in check?
We could work longer hours and maybe even put our machines under more strain, but we can only do that for so long: it would be like trying to run a marathon at sprinting speed.
That doesn’t mean we should abandon all hope.
Instead, to curb price rises, and to lift our living standards over time, we need to improve productivity. Like a marathon runner improving their running technique, we need a way to get faster or better at what we do. A crucial way of doing this is through discovering and using new technology that helps us pump out more, or better quality, goods and services, in a way that can be maintained.
The most influential of these tools (those that have transformed the way we live) are called General-Purpose Technologies – or GPTs for short. The steam engine, cars, electricity and the internet all count as GPTs, because they were widely adopted and became crucial pieces of technology which dramatically yanked up our productivity.
We may not consciously think about it. But imagine what our lives would look like today without electricity, internet and cars. We would be slower, have much less information at our fingertips and would find it harder to work once the sun sets.
As Andrew Leigh points out in his book The Shortest History of Economics, the journey to create the electric bulb itself shows how our productivity has improved. In prehistoric times, producing as much light as a regular household lightbulb using wood fire would have taken our ancestors about 58 hours of foraging for wood. Today, it takes less than a second of work to earn enough to flick a household light bulb on for an hour.
ChatGPT is an example of a tool that could become a general-purpose technology. But the “GPT” in its name actually stands for “generative pre-trained transformer”: a fancy way of saying a piece of software trained using huge amounts of data to offer up human-like answers to questions like mine.
During the pandemic, there was a short-lived surge in the take-up of cloud computing (IT services that businesses can use without owning or running the physical servers, hard drives and networks required themselves). But generally, Australian businesses are behind the curve when it comes to adopting new technologies – and we don’t develop much of it ourselves.
That doesn’t mean we should abandon all hope. Instead, we need to think about the drivers of, and barriers to, adopting technologies such as cloud computing and artificial intelligence: two GPTs in the making.
Kim Nguyen and Jonathan Hambur at the Reserve Bank say these technologies could alter the way we do business. But knowing how to use and make the most of them also requires highly skilled and educated workers.
A website called ChatGPT is raising questions about the role of artificial intelligence in our education, work and relationships.
Nguyen and Hambur’s research involved trawling through the annual reports, job ads and earnings calls of Australian businesses to figure out how much workers’ and managers’ skills matter when it comes to successful adoption of GPTs.
Here’s what they found. Firms which had snagged a board member with experience in the IT industry were 30 percentage points more likely to adopt a GPT. While there were certainly businesses which took up GPT without a technologically skilled board member on their team, these firms generally failed to see much improvement in their profitability after putting a GPT in place.
Basically, having board members with relevant technological experience has been linked to more profitable use of GPT. Of course, the authors point out this could be because firms that appoint technology-savvy board members tend to be more focused on IT in the first place, and therefore more likely to be able to adopt GPT in a way that increases profitability.
But firms with technologically skilled board members were also more likely to look for workers with GPT skills, indicating those workers might also play an important role in profitable GPT adoption. Whatever the exact link, uptake of GPT is linked to higher demand for skilled workers, meaning education and training will be key to nailing the use of these technologies.
While the Reserve Bank’s toolkit is limited to setting interest rates (and, informally, jawboning) the less painful solution to getting inflation under control is to improve our productivity, and therefore the amount of goods and services to go around.
Productivity growth is difficult to measure, and quarter-to-quarter movements can be rocked by things that have little to do with anything. But it has flattened out in recent months, and without productivity growth to match, wages, which have begun to pick up in recent times, will worry the Reserve Bank and may build the case for the Reserve Bank to keep interest rates higher for longer.
ChatGPT has hit the headlines over the past year: from students using it in a bid to boost their marks and to some media companies relying on it to churn out AI-generated content. While it’s yet to join the ranks of coveted general-purpose technologies, ChatGPT is an example of innovation which could turn out to be a game-changer.
Right now, it’s an imperfect tool being put to use by an inexperienced user (me). But I asked ChatGPT if it could write a better opinion piece, and faster than I could. The answer? “I’d love to give it a try!”